normalization for visualizing two samples

Jun Kim

unread,

Jul 1, 2021, 11:36:02 AM7/1/21

to 3D Genomics

Hi,

thank you for the amazing tool.

I have two samples that have similar sequencing depth, but obviously do not have the same exact depth.

I would like to compare them side by side using the same min/max values for color scale.

Is there a way to 'rescale' the two samples so that the visual comparison is more valid?

Thank you,

Jun Kim

Matt Romero

unread,

Jul 13, 2021, 3:37:13 PM7/13/21

to 3D Genomics

Hi Jun Kim,

I had a similar question - did you ever get this answered? This answer would help me as well! I also wonder if normalization for sequencing depth should take place before or after correction? Hope these get answered. Thanks for posting the question!

-Matt

Muhammad Saad Shamim

unread,

Jul 13, 2021, 3:42:11 PM7/13/21

to 3D Genomics

Hey Jun Kim and Matt,

Have you tried using VS mode in Juicebox Desktop? https://aidenlab.gitbook.io/juicebox/comparing-maps

It does attempt to normalize the mean of each dataset and give a fairer comparison, without needing to do any additional post-processing.

Just use the .hic files you already have.

Best,

- Muhammad Saad Shamim

--
You received this message because you are subscribed to the Google Groups "3D Genomics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3d-genomics...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/3d-genomics/5f0c4dd3-e682-45ae-b591-14118e8e587an%40googlegroups.com.

Jun Kim

unread,

Jul 14, 2021, 9:29:30 AM7/14/21

to 3D Genomics

Hi Matt,

Two solutions I used:

1. simply use observed expected (since its relative to expected, the count value of matrix shouldn't matter)

2. convert .hic file to hicexplorer format and normalize two samples based on matrix depth using their function hicNormalize.

The hicexplorer developers suggest normalizing the matrix depth BEFORE correcting via (KR or IC). I think the point was that the correction values and the raw matrix is stored separately, so the correction values computed before depth-normalization would be incorrect to apply in the depth-normalized matrix.

hope that helps,

Jun

Matt Romero

unread,

Jul 19, 2021, 1:51:22 PM7/19/21

to 3D Genomics

Hi Muhammad and Jun,

Thanks so much for the info! I did use the VS mode in juicebox and that seems to work well, I just wanted to do my best to see if I could normalize for read depth because some of my samples are fairly different in this regard.

Jun, funny you mentioned this tool because I was actually in the process of using it when you suggested it. One thing I found that is currently getting fixed by the hicexplorer team is that if you try to normalize and then correct a .cool file using hicCorrectMatrix, something goes wrong. But converting your files from .hic to .h5 can circumvent this problem. Here's my asking naive questions on github (https://github.com/deeptools/HiCExplorer/issues/724).